An approach to mobile robot self-training

Author

Golovko, Vladimir ; Ignatiuk, O. ; Sauta, Vladimir

Author_Institution

Dept. of Comput. & Mech., Brest Polytech. Inst., Byelorussia

fYear

2000

fDate

2000

Firstpage

608

Lastpage

613

Abstract

The unsupervised learning of the autonomous mobile robot is one of the actual research topics. It permits the artificial system to interact successfully with their environment and to avoid obstacles. This paper presents an intelligent control architecture which integrates self-training methods and is available to operate in complex, unknown environment in order to achieve the target. Our approach is based on the reactive obstacle avoidance. The intelligent model integrates different neural networks and permits the robot to perform online learning. The results of experiments are discussed

Keywords

intelligent control; mobile robots; multilayer perceptrons; path planning; unsupervised learning; autonomous mobile robot; intelligent control; multilayer perceptron; neural networks; obstacle avoidance; online learning; self organising; self-training; Artificial intelligence; Artificial neural networks; Intelligent control; Intelligent networks; Intelligent robots; Learning systems; Mobile robots; Neural networks; Robot sensing systems; Unsupervised learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Vehicles Symposium, 2000. IV 2000. Proceedings of the IEEE

Conference_Location

Dearborn, MI

Print_ISBN

0-7803-6363-9

Type

conf

DOI

10.1109/IVS.2000.898415

Filename

898415